Skip to content
skillmaster
FREnglish version : EN
Install

video-understanding

zenstory-ai/video-recap-skills / skills/video-understanding

indexes a video by scene and dialogue , which improves: avoids manually reviewing raw footage

score 9/10high risk581 repo stars
View on GitHub (new tab)

Details

judge’s score
9/10
place
outside the packs
publisher
zenstory-ai
status
new
pinned commit
6b7a816 on GitHub (new tab)
skill’s last commit
Oct 15, 2026
repo stars
581
main type
Video

Judgement

Precise pipeline and triggering, with options, outputs and cache pitfalls. Score capped by the Chinese SKILL.md, the dependency on the MiMo API, and scripts not provided for verification.

reservations

  • The 21 Python scripts (including understand.py, asr.py, vlm.py) are not provided: the announced code cannot be verified.
  • SKILL.md written entirely in Chinese, while the developer works in French.
  • Strong dependency on a specific provider (MiMo API, models mimo-v2.5 and mimo-v2.5-asr) with no alternative described.
  • Context research (research-guide) relies on unframed web searches, with no safeguard on the retrieved content.
sub-scores
Triggering
2/2
Added value
2/2
Precision
2/2
Scope
2/2
Quality
1/2

Alerts

5 alerts, by category

Static scan

  • high risk

    scripts run by the SKILL.md

    scripts/understand.py

  • info risk

    bundled scripts, not run by the SKILL.md

    scripts/agent_text.py, scripts/asr.py, scripts/asr_timing_evidence.py, scripts/briefing/__init__.py, scripts/briefing/builder.py

Network

  • high risk

    Sends extracted images, audio, transcripts and background_research.json to a third party (MiMo) without an explicit warning in the skill.

  • high risk

    The variables MIMO_API_URL, MIMO_ASR_API_URL and MIMO_VIDEO_API_URL can redirect requests and the API key to an arbitrary endpoint.

Instruction

  • medium risk

    Web content retrieved for background_research.json can inject instructions into the VLM context and the narration briefs.

Original description

把视频分析为结构化理解索引：场景检测、ASR 转写、逐场景 VLM 观察、静音窗口、融合时间线和写作 brief。 用于理解、索引或总结视频，也作为后续创作前的分析阶段。输入视频文件；输出 scenes.json、 asr_result.json、vlm_analysis.json、silence_periods.json、timeline_fusion.json、agent_narration_brief.md。 触发词：视频理解、视频分析、视频索引、video understanding、analyze video、看懂视频。

added on Oct 19, 2026 · license MIT